Can Network Structure Predict Cross-Sectional Stock Returns? Evidence from Co-attention Networks in Chin

This study represents the Chinese stock market as co-attention networks based on investors’ correlated searches of stocks in a web portal. We investigate the predictability of network structures on cross-sectional stock returns. Our results show that network structures play different roles in predicting stock returns when peer stocks behave differently. Specifically, an increase in network centrality and network closure among the “historical winner peers” of a focal stock is associated with higher abnormal returns, while such an increase among the “historical loser peers” predicts lower abnormal returns.